Schemas¶

Schema nodes describe structured LM output.

class autoform.Str(*, min=None, max=None, pattern=None)[source]¶

String schema node with optional length and pattern constraints.

Use this node in schema trees passed to autoform.lm_schema_call().

Parameters:

min (int | None) – Optional minimum length of the string.
max (int | None) – Optional maximum length of the string.
pattern (str | None) – Optional regular expression pattern that the string must match.

Example

>>> import autoform as af
>>> name = af.Str(min=1, max=80, pattern=r"^[A-Za-z ]+$")

class autoform.Int(*, min=None, max=None)[source]¶

Integer schema node with optional range constraints.

Use this node in schema trees passed to autoform.lm_schema_call().

Parameters:

Example

>>> import autoform as af
>>> count = af.Int(min=0, max=10)

class autoform.Float(*, min=None, max=None)[source]¶

Number schema node with optional range constraints.

Use this node in schema trees passed to autoform.lm_schema_call().

Parameters:

Example

>>> import autoform as af
>>> score = af.Float(min=0, max=1)

class autoform.Bool[source]¶

Boolean schema node.

Use this node in schema trees passed to autoform.lm_schema_call().

Example

>>> import autoform as af
>>> ok = af.Bool()

class autoform.Enum(*values)[source]¶

Enum schema node with a fixed set of allowed values.

Use this node in schema trees passed to autoform.lm_schema_call().

Parameters:: *values (Any) – Allowed values. Values must be non-empty, share one type, and be JSON scalar values.

Example

>>> import autoform as af
>>> kind = af.Enum("summary", "definition")

class autoform.Doc(text, /)[source]¶

Description node for attaching JSON Schema descriptions.

Use this node in schema trees passed to autoform.lm_schema_call().

Example

>>> import autoform as af
>>> name = af.Str() @ af.Doc("Subject name.")